The present invention relates to heat sinks in general, and more particularly to heat sinks for use in dissipating waste heat generated by electrical or electronic components and assemblies.
Research activities have focused on developing heat sinks to efficiently dissipate heat from highly concentrated heat sources such as microprocessors and computer chips. These heat sources typically have power densities in the range of about 5 to 35 W/cm2 (4 to 31 Btu/ft2s) and relatively small available space for placement of fans, heat exchangers, heat sinks and the like.
At the component level, various types of heat exchangers and heat sinks have been used that apply natural or forced convection or other cooling methods. The most commonly existing heat sinks for microelectronics cooling have generally used air to directly remove heat from the heat source. However, air has a relatively low heat capacity. Such heat sinks are suitable for removing heat from relatively low power heat sources with power density in the range of 5 to 15 W/cm2 (4 to 13 Btu/ft2s). Increases in computing speed resulted in corresponding increases in the power density of the heat sources in the order of 20 to 35 W/cm2 (18 to 31 Btu/ft2s) thus requiring more effective heat sinks. Liquid-cooled heat sinks employing high heat capacity fluids like water and water-glycol solutions are more particularly suited to remove heat from these types of high power density heat sources. One type of liquid cooled heat sink circulates the cooling liquid so that the liquid removes heat from the heat source and is then transferred to a remote location where the heat is easily dissipated into a flowing air stream with the use of a liquid-to-air heat exchanger. These types of heat sinks are characterized as indirect heat sinks.
As computing speeds continue to increase even more dramatically, the corresponding power densities of the devices rise up to 100 W/cm2. The constraints of the necessary cooling system miniaturization coupled with high heat flux calls for extremely efficient, compact, simple and reliable heat sinks such as a thermosiphon. A typical thermosiphon comprises an evaporating section and a condensing section. The heat-generating device is mounted to the evaporating section. In some thermosiphons, the heat-generating device is affixed to the internal surface of the evaporating section where it is submerged in the working fluid. Alternatively, the heat-generating device can also be affixed to the external surface of the evaporating section. The working fluid of the thermosiphon is generally a halocarbon fluid, which circulates in a closed-loop fashion between the evaporating and condensing sections. The captive working fluid changes its state from liquid-to-vapor in the evaporating section as it absorbs heat from the heat-generating device. Reverse transformation of the working fluid from vapor-to-liquid occurs as it rejects heat to a cooling fluid like air flowing on an external finned surface of the condensing section. The thermosiphon relies exclusively on gravity for the motion of the working fluid between the evaporating and condensing sections. As for the motion of the cooling fluid on the external surface of the condensing section, a fluid moving device like an axial fan is employed.
Most electronics devices have high degree of non-uniformity built into them. Thermal management of these devices is subject to two constraints that the thermal engineer must address. First, the heat flux generated by the electronics device is highly non-uniform. Second, the air circulated by the air-moving device like an axial fan is very non-uniformly distributed. Most computer chips have their heat generation concentrated in a very small region in the core of the chip. For example, a typical 40xc3x9740 mm2 computer chip has almost 80% of its total heat flux concentrated in its central 10xc3x9710 mm2 surface. The heat flux distribution in a typical electronics device is shown schematically in FIG. 4. The second non-uniformity is attributed to the attachment of the air-moving device like an axial fan attached to the exterior of the thermosiphon. Axial fan has a large hub which acts as blockage to airflow. The airflow exiting the axial fan is highly concentrated in the peripheral region of the fan blades as shown in FIG. 5. The maximum air velocity is in the tip region of fan blades. The velocity falls off sharply to zero in the central hub region. Under certain flow conditions and blade angle, the local velocity at the root of the fan blade may even become negative, i.e., opposite to the direction of the predominant airflow.
The non-uniformity of airflow is far more pronounced in push mode wherein the fan blows relatively cooler ambient air into the heat exchanger. In pull mode, on the other hand, the fan sucks relatively hotter air from the heat exchanger. For a high heat load push mode is advantageous when airflow rate is low. In order to attain flatter airflow profile entering the heat exchanger face a standoff distance of at least three times the hub diameter is preferable between the fan and the heat exchanger. However, because of packaging constraints only about one-fifth to one-quarter of the hub diameter standoff distance is typically available between the fan and heat exchanger. This is because the airflow at the heat exchanger face is non-uniform.
A limitation of the axial fan relating to smallness of the pressure rise across the fan needs to be borne in mind. The curve of the pressure head developed by the fan falls off very rapidly as the volumetric flow rate of air increases. In other words, the air exiting an axial fan cannot sustain a high-pressure drop through the fins. Therefore, managing the airflow through the heat sink at a low-pressure drop is a very important consideration in the design of a thermosiphon.
It is apparent from the foregoing considerations that from a system""s point of view, the computer chip, heat sink and fan assembly are constrained not only by very non-uniform heat flux but also by non-uniform airflow capable of sustaining small pressure drop across the heat exchanger. Ideally, the airflow should be high in regions of high heat flux and low in regions of low heat flux. Overlaying FIGS. 4 and 5 in push mode clearly reveals that the airflow distribution is opposite to that ideally desired for better heat transfer. This is detrimental to the functioning of a computer chip, as the chip junction temperature becomes high because of inadequate heat removal locally from the core of the chip. The thermal performance penalty attributed to these non-uniformities can be of the order of 25 to 50% compared to the case with uniform heat flux and uniform airflow. Thus thermal solution becomes considerably more challenging when the heat flux as well as the airflow is non-uniform. The difficulty is compounded when the available airflow rate is small. Therefore, careful attention must be paid to the fluid flow and heat transfer boundary conditions when developing the thermal solutions for the computer chips.
The compact thermosiphons intended to fit in a computer case require boiling and condensing processes to occur in close proximity to each other thereby imposing conflicting thermal conditions in a relatively small volume. This poses significant challenges to the process of optimizing the thermosiphon performance.
Thus, what is desired is a thermosiphon optimization process to intensify the processes of boiling, condensation and convective heat transfer at the external surface of the condenser while maintaining low airside pressure drop.
One aspect of the present invention is a heat sink assembly for cooling an electronic device. The heat sink assembly comprises a fan housed in a shroud, the fan having a hub and fan blades extending therefrom for causing an axially directed airflow through the shroud upon rotation of the fan blades. A thermosiphon is positioned at one end of the shroud such that the fan is aligned with the condenser for directing the axial airflow therethrough. The thermosiphon comprises an evaporator defining an evaporating chamber containing a working fluid therein and further including a condenser mounted thereabove. The condenser includes a plurality of tubes forming a tube grouping. Each tube having an opening in fluid communication with the evaporator and for receiving and condensing vapor of the working fluid received from the evaporator. The tubes are axially aligned with the airflow and are laterally positioned such that a lateral width of the tube grouping is approximately equal to a width of the hub and substantially in lateral alignment thereto.
Another aspect of the present invention is a condenser for a heat sink assembly for cooling an electronic device. The heat sink assembly comprises a base having an upper housing affixed thereto wherein the upper housing has open ends. A fan is mounted at one of the open ends, the fan having a hub and fan blades extending therefrom for causing an axially directed airflow through the housing upon rotation of the fan blades. A plurality of tubes is positioned within the housing for transmitting therethrough a vapor of a working fluid. The tubes define a tube grouping such that the tubes are arranged in axial alignment with the fan and laterally positioned such that a lateral width of the tube grouping is approximately equal to a width of the hub and substantially in lateral alignment thereto.
These and other advantages of the invention will be further understood and appreciated by those skilled in the art by reference to the following written specification, claims and appended drawings.